Neurological manifestations in patients with liver failure are characterized by hepatic encephalopathy (HE) and a variable degree of brain edema (BE), the latter localized mainly to the cerebral cortex, but to some extent also to the white matter of the brain.1 Although the molecular and pathophysiological mechanisms underlying these complications remain incompletely understood, there is a growing body of evidence suggesting impairment in whole-body nitrogen metabolism. Indeed, conversion of ammonia to urea by the urea-cycle enzymes is known to be an essential process necessary for the maintenance of health. The failure of such processes is manifested in children by a variety of urea cycle disorders (UCDs). Urea cycle processing is also severely compromised in patients with liver failure who have accumulation of metabolic intermediates. Systemic inflammation, low cerebral perfusion pressure, hypoxia, and metabolic disturbances not involved in elimination of ammonia may act synergistically and aggravate or even result in brain dysfunction.2 Indeed, hyperammonemia results in a syndrome resembling HE with similar clinical manifestations including loss of appetite, vomiting, and hyperventilation initially, and later, lethargy, encephalopathy, seizures, and/or BE. In patients with hyperammonemia and with HE, the level and duration of hyperammonemia influences the reversibility of the cerebral pathology as well as the prognosis.3-5 In this context, it is only natural that removal of ammonia from the circulation (and brain) has become a key target in management of patients with liver failure.
Abbreviations
ALF, acute liver failure; BE, brain edema; CvvHF, continuous venovenous hemofiltration; HE, hepatic encephalopathy; LOLA, L-ornithine-L-aspartate; LOPA, L-ornithine and phenylacetate; UCD, urea cycle disorder; UCL, University College London.
Over the years, many attempts have been made to reduce ammonia by dietary protein restriction, colectomy, exchange transfusion, hemodialysis, peritoneal dialysis, and cross-perfusion with pigs, monkeys, and humans. Indeed, the studies in the older literature may, in light of current knowledge, appear rather bizarre, and the results of these treatments have been by and large less than satisfactory. Administration of lactulose appears to have a beneficial role in patients with subclinical HE6 but has a less convincing effect in patients with overt HE.7 More recently, attempts to remove excess ammonia by albumin dialysis have resulted in improvement in the stage of HE in patients with chronic liver failure,8 but this treatment modality currently remains controversial.9, 10 L-Ornithine-L-aspartate (LOLA) decreases plasma ammonia levels by stimulating synthesis of urea and glutamine and improves HE in patients with cirrhosis. However, LOLA is of little value in patients with minimal HE,11 and a recent study concludes that administration of LOLA in patients with acute liver failure (ALF) did not have any effect on brain dysfunction or on survival.12 This finding is thought to result from the lack of removal of glutamine from the circulation/body, which may instead result in accumulation of glutamine and a rebound rise in the circulating NH4+ concentration. All these observations reported over the last 4-5 decades clearly call for a quantitatively effective method to remove not only NH4+ but also glutamine from the circulation in patients with liver failure. In this context, it may be of value to examine again the information obtained by treating patients with UCD and hyperammonemia.
As early as 1979, Brusilow et al.13 proposed the use of an alternative pathway to reduce the level of hyperammonemia in patients with UCD. The drugs proposed were phenylacetate and sodium benzoate. The former drug combines with the amino acid glutamine to produce phenylacetylglutamine, which can be excreted in the urine. This approach removes the same amount of waste nitrogen as urea, because each mole of phenylacetate removes 2 mol of glutamine. In the ensuing decades, this alternative-pathway therapy became a cornerstone in management of patients with UCD who had acute hyperammonemia with or without addition of continuous venovenous hemofiltration techniques (CvvHF) in order to reduce plasma ammonia levels directly.
In this issue of HEPATOLOGY, Dr. Davies et al.14 from the University College London (UCL) liver failure group in the United Kingdom have adapted the idea of using such alternative pathways to lower circulating ammonia levels via urinary disposal of nitrogenous waste products. They demonstrate a synergy between L-ornithine and phenylacetate (LOPA) in reducing arterial ammonia levels in bile duct–ligated rats to values similar to the control groups. In this model, the lowering of NH4+ appears to be associated with a decrease in the cerebral glutamine/myoinositol ratio and brain water content and an increase in the arterial glutamine concentration and urinary phenylacetylglutamine output. Although these results are quite intriguing and exciting, one important question emerges: Why not simply use phenylacetate to remove excess ammonia in chronic liver failure? As pointed out by the authors,14 the answer is simply that chronic liver failure is most often associated with normal circulating levels of glutamine.15 Under such circumstances, a potentially detrimental effect of therapy with phenylacetate alone, in the absence of ornithine, may be that phenylacetate depletes muscles (and brain) of glutamine and glutamate (i.e., by formation of phenylacetylglutamine), resulting in, or aggravating a state of catabolism due to a concomitant oxidation of branched-chain amino acids (i.e., leucine levels).16 Theoretically, the addition of ornithine to phenylacetate would attenuate depletion of glutamate/glutamine, and thus branched-chain amino acids, when phenylacetate forms phenylacetylglutamine. In addition, the authors demonstrate that administration of phenylacetate without ornithine has no effect on the plasma ammonia level in bile duct–ligated rats. Supported by the data presented here, albeit with a rather short-term perspective, there appears to be a rationale for adding ornithine to phenylacetate to treat hyperammonemia in rats with chronic liver failure. In the setting of ALF, the need for combining ornithine with phenylacetate is less clear, because the plasma concentration of (ornithine and) glutamine is about four times higher than that in controls and patients with cirrhosis.15 Theoretically, administration of phenylacetate alone without ornithine may be sufficient to lower the ammonia concentration and prevent HE and BE in patients with ALF, as is possible in patients with UCD.
Also in this issue of HEPATOLOGY, Dr. Ytrebø and the UCL group17 present another important experimental study in a pig model of ALF demonstrating that LOPA may prevent a rise in plasma and brain ammonia concentrations as well as a rise in intracranial pressure. Indeed, it remains unclear whether administration of phenylacetate alone would have been sufficient, or whether the addition of ornithine is essential. The study design with a study group given both ornithine and phenylacetate and two control groups (sham-operated pigs) and an ALF group (given neither ornithine nor phenylacetate) seems inadequate to answer this question. Replacement of the sham control group with two other groups (a group with ALF given ornithine but no phenylacetate, and a group with ALF given phenylacetate but no ornithine) would have resolved this issue. Although such data collected in a small pilot study (n = 3 in each group), which was initiated before the main data reported here, are given in a supporting table and a supporting figure, they do not explain why arterial glutamine is normalized without also occurring for arterial ammonia, as would have been expected. Certainly, it would be preferable if these metabolic inconsistencies could be clarified in a confirmatory prospective study using phenylacetate before phase 1 clinical studies are initiated in patients with ALF.
Because concomitant renal dysfunction and hyperammonemia is common in patients with liver failure, whether acute or chronic, the use of LOPA has to be explored in this context. CvvHF applied as high-volume hemofiltration (above 35 mL/kg/hour) results in a reasonable clearance of organic acids and ammonia, when used to treat hyperammonemia caused by UCD. However, it remains unknown whether phenylacetyl-glutamine is also cleared during CvvHF therapy in patients with concomitant liver and renal failure.
The lack of stoichiometry between the reduction in arterial ammonia concentration and the increase in urinary phenylacetylglutamine is not accounted for in the two articles.14, 17 The authors speculate that this may result from differences in conjugation pathways, or kinetics of phenylacetylglutamine excretion. Another quantitatively important possibility is that ornithine by itself accelerates the compromised urea cycle in the groups with liver failure.
Even if additional studies convincingly demonstrate that LOPA reduces NH4+ and improves HE and BE in patients with liver failure, another serious concern may be the potential toxicity of this drug. In patients with UCD treated with phenylacetate, severe toxicity with plasma phenylacetate level >6 mmol/L are associated with confusion, lethargy, and emesis. In more severe cases with phenylacetate levels of approximately 10 mmol/L, metabolic acidosis, cerebral edema, hypotension, and cardiovascular collapse and an increased anion gap (up to 52 meq/L) have been reported. Clearly, phase 1 studies in patients with chronic liver failure or ALF should monitor plasma levels of LOPA.
To conclude, we face an exciting new development with adoption of the concept of alternative pathway therapy, in a form that may also secure the metabolic needs in patients with liver failure who have hyperammonemia. The two intriguing studies by the UCL group in this issue of HEPATOLOGY certainly underline the potential for a possible clinical application in management of minimal/low-grade HE as well as for deeper levels of HE in association with both chronic liver failure and ALF. Indeed, we should remember that the development of HE carries an increased risk of complications such as aspiration, pneumonia, sepsis, impaired nutrition, and difficulty in delivering optimum care.18 At present, the only safe option for such patients is admission to a critical care environment and intubation to protect their airway, which is especially necessary in patients with ALF awaiting liver transplantation. If LOPA is going to be used in the clinical setting of HE, we will have to determine and monitor LOPA levels and consider toxicity and functionality during complications such as systemic inflammation and renal failure. That is, the use of LOPA in clinical practice should await studies to determine dosing schedules, especially in cases of renal failure. Another real problem that may need to be overcome is if LOPA is found to improve the level of encephalopathy without improving survival. This relates to the current criteria. For liver transplantation in patients with ALF, which includes development of HE stage 3-4, at least in ALF caused by acetaminophen overdose. In spite of all these concerns, the proposed novel therapeutic intervention with LOPA may truly reduce development of and speed the resolution of HE and provide us with an important step forward in the management paradigm.
